# Mind the Performance Gap: Capability-Behavior Trade-offs in Feature Steering

arXiv: https://arxiv.org/abs/2602.04903 · HTML: https://mlsanity.com/paper/2602.04903/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 20 references that are in the map.
- First posted on arXiv 3 Feb 2026.

## Builds on (most cited first, 17 of 17)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Fine-Tuning Language Models from Human Preferences](https://mlsanity.com/paper/1909.08593/) (2019)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [Universal Adversarial Triggers for Attacking and Analyzing NLP](https://mlsanity.com/paper/1908.07125/) (2019)
- [Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task](https://mlsanity.com/paper/2210.13382/) (2022)
- [Emergent Linear Representations in World Models of Self-Supervised Sequence Models](https://mlsanity.com/paper/2309.00941/) (2023)
- [Extracting Latent Steering Vectors from Pretrained Language Models](https://mlsanity.com/paper/2205.05124/) (2022)
- [InFoBench: Evaluating Instruction Following Ability in Large Language Models](https://mlsanity.com/paper/2401.03601/) (2024)
- [Steering Language Model Refusal with Sparse Autoencoders](https://mlsanity.com/paper/2411.11296/) (2024)
- [AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint](https://mlsanity.com/paper/2506.07022/) (2025)
- [A Unified Understanding and Evaluation of Steering Methods](https://mlsanity.com/paper/2502.02716/) (2025)
- [Towards Reliable Evaluation of Behavior Steering Interventions in LLMs](https://mlsanity.com/paper/2410.17245/) (2024)
- [Towards Unifying Interpretability and Control: Evaluation via Intervention](https://mlsanity.com/paper/2411.04430/) (2024)
- [Sparse Neurons Carry Strong Signals of Question Ambiguity in LLMs](https://mlsanity.com/paper/2509.13664/) (2025)

## Built on by (most cited first, 3 of 3)

- [WASD: Locating Critical Neurons as Sufficient Conditions for Explaining and Controlling LLM Behavior](https://mlsanity.com/paper/2603.18474/) (2026)
- [Does Steering Break Your Model? A Multi-Dimensional Evaluation Suite for LLM Steering Methods](https://mlsanity.com/paper/2610.07722/) (2026)
- [Steering Under Compression: Dose-Response, Capability Cost, and Failure Asymmetry in Quantized LLMs](https://mlsanity.com/paper/2609.06473/) (2026)

## Abstract (arXiv)

Feature steering has emerged as a promising approach for controlling LLM behavior through direct manipulation of internal representations, offering advantages over prompt engineering. However, its practical effectiveness in real-world applications remains poorly understood, particularly regarding potential trade-offs with output quality. We show that feature steering methods substantially degrade model performance even when successfully controlling target behaviors, a critical trade-off. Specifically, we evaluate Goodfire's Auto Steer against prompt engineering baselines across 14 steering queries (covering innocuous and safety-relevant behaviors) on 171 Massive Multitask Language Understanding (MMLU) questions using Llama-8B and Llama-70B, measuring accuracy, coherence, and behavioral control. Our findings show that Auto Steer successfully modifies target behaviors (achieving scores of 3.33 vs. 2.98 for prompting on Llama-8B and 3.57 vs. 3.10 on Llama-70B), but causes dramatic performance degradation: accuracy on the MMLU questions drops from 66% to 46% on Llama-8B and 87% to 73% on Llama-70B, with coherence falling from 4.62 to 2.24 and 4.94 to 3.89 respectively. Simple prompting achieves the best overall balance. These findings highlight limitations of current feature steering methods for practical deployment where task performance cannot be sacrificed. More broadly, our work demonstrates that mechanistic control methods face fundamental capability-behavior trade-offs that must be empirically characterized before deployment.